updating to fused (#2293)
This commit is contained in:
@@ -49,7 +49,7 @@ class TestPackedLlama(unittest.TestCase):
|
||||
"gradient_accumulation_steps": 4,
|
||||
"output_dir": temp_dir,
|
||||
"learning_rate": 0.00001,
|
||||
"optimizer": "adamw_torch",
|
||||
"optimizer": "adamw_torch_fused",
|
||||
"lr_scheduler": "cosine",
|
||||
"max_steps": 5,
|
||||
"use_tensorboard": True,
|
||||
|
||||
Reference in New Issue
Block a user